Eight OpenAI Lean Formalizations Found Trivially Provable with False Proofs

Blogger Timeroot found that at least eight comparator challenges in OpenAI's Lean formalization repository can be trivially passed with false proofs due to missing constraints. Elliot Glazer used Claude Opus to analyze the issue, while Kyle Cranmer clarified the problem lies in the formalizations themselves, not the proof system.

2026-10-09 ~ 2026-10-09 · 3 related posts

• 8 of OpenAI's Lean formalization challenges are broken and trivially hackable — gklambauer · 2026-10-09
• 8 Lean formalizations in OpenAI's repo called "broken"; Opus 5.5 weighs in on Comparator critiques — ctjlewis · 2026-10-09
• 8 Lean Formalizations in OpenAI's Repo 'Broken' — Wrong Proofs Pass Verification — AlexTensor · 2026-10-09